Ajou University to lead national project tackling AI data center heat

Ajou University announced Thursday that it has been selected consecutively as the lead research institution for large-scale national R&D and workforce training projects promoted by the Ministry of Trade, Industry and Energy and the Ministry of Climate, Environment and Energy.

Through the selection, Ajou University will lead national R&D efforts and the training of specialized talent in AI data center thermal management — a field that has emerged as a critical challenge alongside the rapid advancement of AI.

The total project budget is approximately 38.8 billion won ($26.5 million).

AI data center power and heat management

Ajou University has been selected as the lead institution for the "Heat Management Technology Development and Demonstration for Data Center Carbon Neutrality" project, a Ministry of Trade, Industry and Energy initiative overseen by the Korea Evaluation Institute of Industrial Technology (KEIT).

The project aims to develop next-generation thermal management technologies that efficiently cool data centers — where power consumption has surged with the spread of AI — to reduce energy use and cut carbon emissions. Running from April 2026 through December 2030, the 4-year, 9-month initiative will receive total research funding of 35.4 billion won, of which 27.6 billion won comes from government support.

Data centers house vast numbers of computer servers that generate enormous heat, and the cooling systems required to manage that heat consume nearly half of a facility's total power. More than 95 percent of domestic data centers currently rely on air-cooling systems, which are relatively inefficient.

Researchers plan to reduce data center Power Usage Effectiveness (PUE) from the current level of around 1.5 to below 1.2, with a further goal of reaching PUE 1.1 by integrating next-generation liquid-cooling technology. PUE measures a data center's energy efficiency by dividing total power consumption by the power used by IT equipment alone — the remainder going largely to cooling and backup power systems. A value closer to 1 indicates less wasted energy, with the decimal portion reflecting primarily the share of power devoted to cooling. High-performance AI servers such as those using GPUs generate particularly intense heat, and the inability to cool them efficiently translates directly into steep electricity costs, driving active research and investment in the field.

The project is divided into three sub-tasks carried out under the supervision of Ajou University professor Lee Jeong-ho through industry-academia-research collaboration.

The first sub-task, led by Ajou University, involves developing "multi-scale distributed temperature sensing (DTS)" technology using optical fiber to measure temperatures in real time across data center components — chips, servers and racks — along with an AI-based temperature-load prediction algorithm.

The second sub-task, led by CoolMicro, focuses on developing a low-power slim-fan-based in-rack cooling system that dissipates heat directly at the source, as well as an on-device AI technology capable of autonomous control decisions.

The third sub-task, led by the Korea Institute of Machinery and Materials, involves building an integrated monitoring platform that provides a comprehensive view of thermal management conditions across an entire data center, with performance to be verified at live operating facilities including those run by Lotte Innovate.

Training master's and doctoral talent in AI, defense and thermal-fluid systems — 3.4 billion won committed

Ajou University has also been designated as the lead institution for the "Smart Energy-AI Thermal-Fluid Equipment Systems Mid-Sized Company Specialized Workforce Training Project," a Ministry of Climate, Environment and Energy initiative promoted by the Korea Energy Technology Evaluation and Planning Institute (KETEP).

Running from April 2026 through December 2030, the 4-year, 9-month project will invest a total of 3.39 billion won to train master's- and doctoral-level specialists ready for immediate deployment in AI data centers, the defense industry, and smart thermal-fluid systems — covering areas such as air conditioning, pumps and heat exchangers.

Ajou University has formed a consortium with universities including Chung-Ang University and KAIST, as well as major companies including SK Innovation, Hanon Systems and Firstec. The university is also partnering with 14 collaborating organizations — including Hanwha Aerospace, LG Electronics, LIG Nex1 and Doosan Enerbility — to deliver industry-tailored education.

The university plans to establish three specialized education tracks covering AI data center thermal management, AI defense industry thermal management, and AI smart thermal-fluid systems. It will also run project-based learning centered on solving real corporate challenges, along with field training and internship programs to strengthen students' practical skills.

Lee Jeong-ho, a professor in Ajou University's Department of Mechanical Engineering and the project's principal investigator, said data centers are being built at a rapid pace as the full-scale AI era arrives, yet effective solutions to power supply challenges have yet to emerge.

Lee added that the university aims to become a first mover in the AI data center field through the development of innovative thermal management technologies and the cultivation of specialized talent.